geant4-data: symlink only specific data dirs (#47367)

Currently, the `geant4-data` spec creates symlink to all of its
dependencies, and it does so by globbing their `share/` directories.
This works very well for the way Spack installs these, but it doesn't
work for anybody wanting to use e.g. the Geant4 data on CVMFS. See pull
request #47298. This commit changes the way the `geant4-data` spec
works. It no longer blindly globs directories and makes symlinks, but it
asks its dependencies specifically for the name of their data directory.
This should allow us to use Spack to use the CVMFS installations as
externals.

16 files changed, 113 insertions, 37 deletions

diff --git a/var/spack/repos/builtin/packages/g4abla/package.py b/var/spack/repos/builtin/packages/g4abla/package.py
index 710d8de011..c36cb6f148 100644
--- a/var/spack/repos/builtin/packages/g4abla/package.py
+++ b/var/spack/repos/builtin/packages/g4abla/package.py
@@ -24,13 +24,18 @@ class G4abla(Package):
 
     def install(self, spec, prefix):
         mkdirp(join_path(prefix.share, "data"))
-        install_path = join_path(prefix.share, "data", "G4ABLA{0}".format(self.version))
+        install_path = join_path(prefix.share, "data", self.g4datasetname)
         install_tree(self.stage.source_path, install_path)
 
     def setup_dependent_run_environment(self, env, dependent_spec):
-        install_path = join_path(self.prefix.share, "data", "G4ABLA{0}".format(self.version))
+        install_path = join_path(self.prefix.share, "data", self.g4datasetname)
         env.set("G4ABLADATA", install_path)
 
     def url_for_version(self, version):
         """Handle version string."""
         return "http://geant4-data.web.cern.ch/geant4-data/datasets/G4ABLA.%s.tar.gz" % version
+
+    @property
+    def g4datasetname(self):
+        spec = self.spec
+        return "G4ABLA{0}".format(spec.version)
diff --git a/var/spack/repos/builtin/packages/g4emlow/package.py b/var/spack/repos/builtin/packages/g4emlow/package.py
index 87c25b950b..458f6da76b 100644
--- a/var/spack/repos/builtin/packages/g4emlow/package.py
+++ b/var/spack/repos/builtin/packages/g4emlow/package.py
@@ -35,13 +35,18 @@ class G4emlow(Package):
 
     def install(self, spec, prefix):
         mkdirp(join_path(prefix.share, "data"))
-        install_path = join_path(prefix.share, "data", "G4EMLOW{0}".format(self.version))
+        install_path = join_path(prefix.share, "data", self.g4datasetname)
         install_tree(self.stage.source_path, install_path)
 
     def setup_dependent_run_environment(self, env, dependent_spec):
-        install_path = join_path(self.prefix.share, "data", "G4EMLOW{0}".format(self.version))
+        install_path = join_path(self.prefix.share, "data", self.g4datasetname)
         env.set("G4LEDATA", install_path)
 
     def url_for_version(self, version):
         """Handle version string."""
         return "https://geant4-data.web.cern.ch/geant4-data/datasets/G4EMLOW.%s.tar.gz" % version
+
+    @property
+    def g4datasetname(self):
+        spec = self.spec
+        return "G4EMLOW{0}".format(spec.version)
diff --git a/var/spack/repos/builtin/packages/g4ensdfstate/package.py b/var/spack/repos/builtin/packages/g4ensdfstate/package.py
index 59fe04f45d..6cb3904756 100644
--- a/var/spack/repos/builtin/packages/g4ensdfstate/package.py
+++ b/var/spack/repos/builtin/packages/g4ensdfstate/package.py
@@ -25,11 +25,11 @@ class G4ensdfstate(Package):
 
     def install(self, spec, prefix):
         mkdirp(join_path(prefix.share, "data"))
-        install_path = join_path(prefix.share, "data", "G4ENSDFSTATE{0}".format(self.version))
+        install_path = join_path(prefix.share, "data", self.g4datasetname)
         install_tree(self.stage.source_path, install_path)
 
     def setup_dependent_run_environment(self, env, dependent_spec):
-        install_path = join_path(self.prefix.share, "data", "G4ENSDFSTATE{0}".format(self.version))
+        install_path = join_path(self.prefix.share, "data", self.g4datasetname)
         env.set("G4ENSDFSTATEDATA", install_path)
 
     def url_for_version(self, version):
@@ -37,3 +37,8 @@ class G4ensdfstate(Package):
         return (
             "http://geant4-data.web.cern.ch/geant4-data/datasets/G4ENSDFSTATE.%s.tar.gz" % version
         )
+
+    @property
+    def g4datasetname(self):
+        spec = self.spec
+        return "G4ENSDFSTATE{0}".format(spec.version)
diff --git a/var/spack/repos/builtin/packages/g4incl/package.py b/var/spack/repos/builtin/packages/g4incl/package.py
index 479ce2dda2..056910d714 100644
--- a/var/spack/repos/builtin/packages/g4incl/package.py
+++ b/var/spack/repos/builtin/packages/g4incl/package.py
@@ -25,13 +25,18 @@ class G4incl(Package):
 
     def install(self, spec, prefix):
         mkdirp(join_path(prefix.share, "data"))
-        install_path = join_path(prefix.share, "data", "G4INCL{0}".format(self.version))
+        install_path = join_path(prefix.share, "data", self.g4datasetname)
         install_tree(self.stage.source_path, install_path)
 
     def setup_dependent_run_environment(self, env, dependent_spec):
-        install_path = join_path(self.prefix.share, "data", "G4INCL{0}".format(self.version))
+        install_path = join_path(self.prefix.share, "data", self.g4datasetname)
         env.set("G4INCLDATA", install_path)
 
     def url_for_version(self, version):
         """Handle version string."""
         return "http://geant4-data.web.cern.ch/geant4-data/datasets/G4INCL.%s.tar.gz" % version
+
+    @property
+    def g4datasetname(self):
+        spec = self.spec
+        return "G4INCL{0}".format(spec.version)
diff --git a/var/spack/repos/builtin/packages/g4ndl/package.py b/var/spack/repos/builtin/packages/g4ndl/package.py
index a8c8f68898..fa94da01b3 100644
--- a/var/spack/repos/builtin/packages/g4ndl/package.py
+++ b/var/spack/repos/builtin/packages/g4ndl/package.py
@@ -25,13 +25,18 @@ class G4ndl(Package):
 
     def install(self, spec, prefix):
         mkdirp(join_path(prefix.share, "data"))
-        install_path = join_path(prefix.share, "data", "G4NDL{0}".format(self.version))
+        install_path = join_path(prefix.share, "data", self.g4datasetname)
         install_tree(self.stage.source_path, install_path)
 
     def setup_dependent_run_environment(self, env, dependent_spec):
-        install_path = join_path(self.prefix.share, "data", "G4NDL{0}".format(self.version))
+        install_path = join_path(self.prefix.share, "data", self.g4datasetname)
         env.set("G4NEUTRONHPDATA", install_path)
 
     def url_for_version(self, version):
         """Handle version string."""
         return "http://geant4-data.web.cern.ch/geant4-data/datasets/G4NDL.%s.tar.gz" % version
+
+    @property
+    def g4datasetname(self):
+        spec = self.spec
+        return "G4NDL{0}".format(spec.version)
diff --git a/var/spack/repos/builtin/packages/g4neutronxs/package.py b/var/spack/repos/builtin/packages/g4neutronxs/package.py
index 39f6915346..595d86c71e 100644
--- a/var/spack/repos/builtin/packages/g4neutronxs/package.py
+++ b/var/spack/repos/builtin/packages/g4neutronxs/package.py
@@ -24,11 +24,11 @@ class G4neutronxs(Package):
 
     def install(self, spec, prefix):
         mkdirp(join_path(prefix.share, "data"))
-        install_path = join_path(prefix.share, "data", "G4NEUTRONXS{0}".format(self.version))
+        install_path = join_path(prefix.share, "data", self.g4datasetname)
         install_tree(self.stage.source_path, install_path)
 
     def setup_dependent_run_environment(self, env, dependent_spec):
-        install_path = join_path(self.prefix.share, "data", "G4NEUTRONXS{0}".format(self.version))
+        install_path = join_path(self.prefix.share, "data", self.g4datasetname)
         env.set("G4NEUTRONXSDATA", install_path)
 
     def url_for_version(self, version):
@@ -36,3 +36,8 @@ class G4neutronxs(Package):
         return (
             "http://geant4-data.web.cern.ch/geant4-data/datasets/G4NEUTRONXS.%s.tar.gz" % version
         )
+
+    @property
+    def g4datasetname(self):
+        spec = self.spec
+        return "G4NEUTRONXS{0}".format(spec.version)
diff --git a/var/spack/repos/builtin/packages/g4nudexlib/package.py b/var/spack/repos/builtin/packages/g4nudexlib/package.py
index 073aebcc7f..2e02321fe4 100644
--- a/var/spack/repos/builtin/packages/g4nudexlib/package.py
+++ b/var/spack/repos/builtin/packages/g4nudexlib/package.py
@@ -23,13 +23,18 @@ class G4nudexlib(Package):
 
     def install(self, spec, prefix):
         mkdirp(join_path(prefix.share, "data"))
-        install_path = join_path(prefix.share, "data", "G4NUDEXLIB{0}".format(self.version))
+        install_path = join_path(prefix.share, "data", self.g4datasetname)
         install_tree(self.stage.source_path, install_path)
 
     def setup_dependent_run_environment(self, env, dependent_spec):
-        install_path = join_path(self.prefix.share, "data", "G4NUDEXLIB{0}".format(self.version))
+        install_path = join_path(self.prefix.share, "data", self.g4datasetname)
         env.set("G4NUDEXLIBDATA", install_path)
 
     def url_for_version(self, version):
         """Handle version string."""
         return "http://geant4-data.web.cern.ch/geant4-data/datasets/G4NUDEXLIB.%s.tar.gz" % version
+
+    @property
+    def g4datasetname(self):
+        spec = self.spec
+        return "G4NUDEXLIB{0}".format(spec.version)
diff --git a/var/spack/repos/builtin/packages/g4particlexs/package.py b/var/spack/repos/builtin/packages/g4particlexs/package.py
index fb86575b46..209831dfa6 100644
--- a/var/spack/repos/builtin/packages/g4particlexs/package.py
+++ b/var/spack/repos/builtin/packages/g4particlexs/package.py
@@ -28,11 +28,11 @@ class G4particlexs(Package):
 
     def install(self, spec, prefix):
         mkdirp(join_path(prefix.share, "data"))
-        install_path = join_path(prefix.share, "data", "G4PARTICLEXS{0}".format(self.version))
+        install_path = join_path(prefix.share, "data", self.g4datasetname)
         install_tree(self.stage.source_path, install_path)
 
     def setup_dependent_run_environment(self, env, dependent_spec):
-        install_path = join_path(self.prefix.share, "data", "G4PARTICLEXS{0}".format(self.version))
+        install_path = join_path(self.prefix.share, "data", self.g4datasetname)
         env.set("G4PARTICLEXSDATA", install_path)
 
     def url_for_version(self, version):
@@ -40,3 +40,8 @@ class G4particlexs(Package):
         return (
             "http://geant4-data.web.cern.ch/geant4-data/datasets/G4PARTICLEXS.%s.tar.gz" % version
         )
+
+    @property
+    def g4datasetname(self):
+        spec = self.spec
+        return "G4PARTICLEXS{0}".format(spec.version)
diff --git a/var/spack/repos/builtin/packages/g4photonevaporation/package.py b/var/spack/repos/builtin/packages/g4photonevaporation/package.py
index 1c847dec3e..099ded6a5f 100644
--- a/var/spack/repos/builtin/packages/g4photonevaporation/package.py
+++ b/var/spack/repos/builtin/packages/g4photonevaporation/package.py
@@ -27,13 +27,11 @@ class G4photonevaporation(Package):
 
     def install(self, spec, prefix):
         mkdirp(join_path(prefix.share, "data"))
-        install_path = join_path(prefix.share, "data", "PhotonEvaporation{0}".format(self.version))
+        install_path = join_path(prefix.share, "data", self.g4datasetname)
         install_tree(self.stage.source_path, install_path)
 
     def setup_dependent_run_environment(self, env, dependent_spec):
-        install_path = join_path(
-            self.prefix.share, "data", "PhotonEvaporation{0}".format(self.version)
-        )
+        install_path = join_path(self.prefix.share, "data", self.g4datasetname)
         env.set("G4LEVELGAMMADATA", install_path)
 
     def url_for_version(self, version):
@@ -42,3 +40,8 @@ class G4photonevaporation(Package):
             "http://geant4-data.web.cern.ch/geant4-data/datasets/G4PhotonEvaporation.%s.tar.gz"
             % version
         )
+
+    @property
+    def g4datasetname(self):
+        spec = self.spec
+        return "PhotonEvaporation{0}".format(spec.version)
diff --git a/var/spack/repos/builtin/packages/g4pii/package.py b/var/spack/repos/builtin/packages/g4pii/package.py
index a1aa5590bc..c9cf78e5cb 100644
--- a/var/spack/repos/builtin/packages/g4pii/package.py
+++ b/var/spack/repos/builtin/packages/g4pii/package.py
@@ -22,13 +22,18 @@ class G4pii(Package):
 
     def install(self, spec, prefix):
         mkdirp(join_path(prefix.share, "data"))
-        install_path = join_path(prefix.share, "data", "G4PII{0}".format(self.version))
+        install_path = join_path(prefix.share, "data", self.g4datasetname)
         install_tree(self.stage.source_path, install_path)
 
     def setup_dependent_run_environment(self, env, dependent_spec):
-        install_path = join_path(self.prefix.share, "data", "G4PII{0}".format(self.version))
+        install_path = join_path(self.prefix.share, "data", self.g4datasetname)
         env.set("G4PIIDATA", install_path)
 
     def url_for_version(self, version):
         """Handle version string."""
         return "https://geant4-data.web.cern.ch/geant4-data/datasets/G4PII.1.3.tar.gz" % version
+
+    @property
+    def g4datasetname(self):
+        spec = self.spec
+        return "G4PII{0}".format(spec.version)
diff --git a/var/spack/repos/builtin/packages/g4radioactivedecay/package.py b/var/spack/repos/builtin/packages/g4radioactivedecay/package.py
index 69f5681b64..33a75291ff 100644
--- a/var/spack/repos/builtin/packages/g4radioactivedecay/package.py
+++ b/var/spack/repos/builtin/packages/g4radioactivedecay/package.py
@@ -27,13 +27,11 @@ class G4radioactivedecay(Package):
 
     def install(self, spec, prefix):
         mkdirp(join_path(prefix.share, "data"))
-        install_path = join_path(prefix.share, "data", "RadioactiveDecay{0}".format(self.version))
+        install_path = join_path(prefix.share, "data", self.g4datasetname)
         install_tree(self.stage.source_path, install_path)
 
     def setup_dependent_run_environment(self, env, dependent_spec):
-        install_path = join_path(
-            self.prefix.share, "data", "RadioactiveDecay{0}".format(self.version)
-        )
+        install_path = join_path(self.prefix.share, "data", self.g4datasetname)
         env.set("G4RADIOACTIVEDATA", install_path)
 
     def url_for_version(self, version):
@@ -42,3 +40,8 @@ class G4radioactivedecay(Package):
             "http://geant4-data.web.cern.ch/geant4-data/datasets/G4RadioactiveDecay.%s.tar.gz"
             % version
         )
+
+    @property
+    def g4datasetname(self):
+        spec = self.spec
+        return "RadioactiveDecay{0}".format(spec.version)
diff --git a/var/spack/repos/builtin/packages/g4realsurface/package.py b/var/spack/repos/builtin/packages/g4realsurface/package.py
index 28f335bba0..1d92494364 100644
--- a/var/spack/repos/builtin/packages/g4realsurface/package.py
+++ b/var/spack/repos/builtin/packages/g4realsurface/package.py
@@ -25,11 +25,11 @@ class G4realsurface(Package):
 
     def install(self, spec, prefix):
         mkdirp(join_path(prefix.share, "data"))
-        install_path = join_path(prefix.share, "data", "RealSurface{0}".format(self.version))
+        install_path = join_path(prefix.share, "data", self.g4datasetname)
         install_tree(self.stage.source_path, install_path)
 
     def setup_dependent_run_environment(self, env, dependent_spec):
-        install_path = join_path(self.prefix.share, "data", "RealSurface{0}".format(self.version))
+        install_path = join_path(self.prefix.share, "data", self.g4datasetname)
         env.set("G4REALSURFACEDATA", install_path)
 
     def url_for_version(self, version):
@@ -39,3 +39,8 @@ class G4realsurface(Package):
                 "G4" if version > Version("1.0") else "", version
             )
         )
+
+    @property
+    def g4datasetname(self):
+        spec = self.spec
+        return "RealSurface{0}".format(spec.version)
diff --git a/var/spack/repos/builtin/packages/g4saiddata/package.py b/var/spack/repos/builtin/packages/g4saiddata/package.py
index 371811381b..b4da318580 100644
--- a/var/spack/repos/builtin/packages/g4saiddata/package.py
+++ b/var/spack/repos/builtin/packages/g4saiddata/package.py
@@ -23,13 +23,18 @@ class G4saiddata(Package):
 
     def install(self, spec, prefix):
         mkdirp(join_path(prefix.share, "data"))
-        install_path = join_path(prefix.share, "data", "G4SAIDDATA{0}".format(self.version))
+        install_path = join_path(prefix.share, "data", self.g4datasetname)
         install_tree(self.stage.source_path, install_path)
 
     def setup_dependent_run_environment(self, env, dependent_spec):
-        install_path = join_path(self.prefix.share, "data", "G4SAIDDATA{0}".format(self.version))
+        install_path = join_path(self.prefix.share, "data", self.g4datasetname)
         env.set("G4SAIDXSDATA", install_path)
 
     def url_for_version(self, version):
         """Handle version string."""
         return "http://geant4-data.web.cern.ch/geant4-data/datasets/G4SAIDDATA.%s.tar.gz" % version
+
+    @property
+    def g4datasetname(self):
+        spec = self.spec
+        return "G4SAIDDATA{0}".format(spec.version)
diff --git a/var/spack/repos/builtin/packages/g4tendl/package.py b/var/spack/repos/builtin/packages/g4tendl/package.py
index 92c12423c2..1a2e909511 100644
--- a/var/spack/repos/builtin/packages/g4tendl/package.py
+++ b/var/spack/repos/builtin/packages/g4tendl/package.py
@@ -24,13 +24,18 @@ class G4tendl(Package):
 
     def install(self, spec, prefix):
         mkdirp(join_path(prefix.share, "data"))
-        install_path = join_path(prefix.share, "data", "G4TENDL{0}".format(self.version))
+        install_path = join_path(prefix.share, "data", self.g4datasetname)
         install_tree(self.stage.source_path, install_path)
 
     def setup_dependent_run_environment(self, env, dependent_spec):
-        install_path = join_path(self.prefix.share, "data", "G4TENDL{0}".format(self.version))
+        install_path = join_path(self.prefix.share, "data", self.g4datasetname)
         env.set("G4PARTICLEHPDATA", install_path)
 
     def url_for_version(self, version):
         """Handle version string."""
         return "http://geant4-data.web.cern.ch/geant4-data/datasets/G4TENDL.%s.tar.gz" % version
+
+    @property
+    def g4datasetname(self):
+        spec = self.spec
+        return "G4TENDL{0}".format(spec.version)
diff --git a/var/spack/repos/builtin/packages/g4urrpt/package.py b/var/spack/repos/builtin/packages/g4urrpt/package.py
index e9ff5df284..a04e6d5d1a 100644
--- a/var/spack/repos/builtin/packages/g4urrpt/package.py
+++ b/var/spack/repos/builtin/packages/g4urrpt/package.py
@@ -23,13 +23,18 @@ class G4urrpt(Package):
 
     def install(self, spec, prefix):
         mkdirp(join_path(prefix.share, "data"))
-        install_path = join_path(prefix.share, "data", "G4URRPT{0}".format(self.version))
+        install_path = join_path(prefix.share, "data", self.g4datasetname)
         install_tree(self.stage.source_path, install_path)
 
     def setup_dependent_run_environment(self, env, dependent_spec):
-        install_path = join_path(self.prefix.share, "data", "G4URRPT{0}".format(self.version))
+        install_path = join_path(self.prefix.share, "data", self.g4datasetname)
         env.set("G4URRPTDATA", install_path)
 
     def url_for_version(self, version):
         """Handle version string."""
         return "http://geant4-data.web.cern.ch/geant4-data/datasets/G4URRPT.%s.tar.gz" % version
+
+    @property
+    def g4datasetname(self):
+        spec = self.spec
+        return "G4URRPT{0}".format(spec.version)
diff --git a/var/spack/repos/builtin/packages/geant4-data/package.py b/var/spack/repos/builtin/packages/geant4-data/package.py
index 201d2e8d88..3b934da392 100644
--- a/var/spack/repos/builtin/packages/geant4-data/package.py
+++ b/var/spack/repos/builtin/packages/geant4-data/package.py
@@ -3,7 +3,6 @@
 #
 # SPDX-License-Identifier: (Apache-2.0 OR MIT)
 
-import glob
 import os
 
 from spack.package import *
@@ -204,5 +203,11 @@ class Geant4Data(BundlePackage):
     def install(self, spec, prefix):
         with working_dir(self.datadir, create=True):
             for s in spec.dependencies():
-                for d in glob.glob("{0}/data/*".format(s.prefix.share)):
-                    os.symlink(d, os.path.basename(d))
+                if not s.name.startswith("g4"):
+                    continue
+
+                if not hasattr(s.package, "g4datasetname"):
+                    raise InstallError(f"Dependency `{s.name}` does not expose `g4datasetname`")
+
+                d = "{0}/data/{1}".format(s.prefix.share, s.package.g4datasetname)
+                os.symlink(d, os.path.basename(d))